Introduction

In polar regions, benthic communities in shallow areas contribute up to 83% of total productivity (Glud 2008) and play an important role in carbon cycling (Krause-Jensen and Duarte 2016). In the Arctic, and especially in Greenland, glacial ice is discharged and lost at an increasing rate (Howat et al. 2007; Kjeldsen et al. 2015). The released icebergs increase the risk of ice scouring while melt water is likely to impact fjord circulations, surface-water temperature and salinity (Mortensen et al. 2011, 2014). These physical parameters are often suggested as important drivers of benthic recruitment in coastal areas (Wiencke et al. 2004; Fredersdorf et al. 2009; Woodson et al. 2012; Yakovis et al. 2013) which can influence benthic productivity. For example, in the sub-Arctic White Sea, the mean seasonal air temperatures were found to explain 20% of the annual variation in barnacle and ascidian recruitment (Yakovis et al. 2013). Temperature has also been found to influence the germination success of kelp zoospores (Fredersdorf et al. 2009). In addition, ice scouring can cause severe local damage to benthic communities and recolonization of newly scoured areas depends on the settlement rate of recruiting and migratory organisms (Gutt 2001).

Biotic interactions can also be key drivers of benthic recruitment (Bertness and Callaway 1994). For example, adult intertidal barnacles may attract barnacle larvae by chemical cues (Jarrett 1997). Nevertheless, the drivers of recruitment patterns are still poorly understood due to complex lifecycles, involving pelagic larval or spore stages before settling in the benthic environment. Due to the mosaic of multiple abiotic and biotic drivers possibly acting at different spatial and temporal scales, settlement and recruitment of marine benthos are often highly variable in time and space (Watson and Barnes 2004).

Few have studied settlement and recruitment patterns of subtidal benthos in Arctic fjords (Fetzer et al. 2002; Beuchel and Gulliksen 2008; Yakovis et al. 2013), and there are no studies from Greenland, which holds 12% of the world’s coastline. This study is a first attempt to investigate potential drivers of subtidal macrobenthic settlement and recruitment across a temporal and spatial scale in Greenland. We deployed settling plates, a standard tool in quantifying recruitment and early growth (Khalaman et al. 2016), at varying distances from a glacial output in a sub-Arctic fjord system in south west Greenland, while monitoring key physical parameters such as temperature, light and salinity.

Materials and methods

The study was conducted in the sub-Arctic Godthåbsfjord system, south west Greenland (64°N, 51°W) (Fig. 1), covering an area of 2013 km2 where the inner part of the main fjord is in contact with three tidal outlet glaciers (Mortensen et al. 2011). The system is influenced by strong tides with a neap spring tidal range between 0.81 and 5.36 m (Mortensen et al. 2011) and during spring and summer, it receives large amounts of freshwater from melting glaciers and runoff from snow-covered areas (Van As et al. 2014). In winter, the innermost part of the main fjord and some of the fjord branches, e.g. Kobbefjord (K), are covered with sea ice. This results in a system characterized by large seasonal variation in key physical parameters such as light, temperature and salinity. From August 2009 to August 2010, 18 settling plates (11 × 11 cm = 0.0121 m2) were deployed at three stations in the Godthåbsfjord system (Q1, Q2, Q3) (Fig. 1), with 6 plates per station. In April 2016, 80 settling plates (12 × 16 cm = 0.0192 m2) were deployed across 7 stations in the Godthåbsfjord system (G, F, K1a, K1b, K1c, K2, K3) (Fig. 1). Half of the plates in each station were harvested after 4 months (Apr 2016–Aug 2016) and replaced with new plates (deployed then from Aug 2016–Apr 2017) while the remaining stayed at the location for a full year (Apr 2016–Apr 2017). Per station and per time period, we had 4 replicates. The deployment covered three spatial scales: three fjord branches (stations G, F, K) (> 30 km apart), three sites within a single fjord (Kobbefjord; K1, K2, K3) (2 km apart) and three stations within a site in Kobbefjord (K1a, K1b, K1c) (20 m apart) (Fig. 1). In 2009, we used PVC settling plates (transparent), treated with coarse sandpaper (Grit 24) to create a rough surface. In 2016, we used polyethylene cutting boards (white) also treated with coarse sandpaper. The plates were attached with cable ties to a larger PVC baseplate (grey) (30 × 40 cm in 2009) (40 × 60 cm in 2016). All the plates were positioned horizontally in the water column. Deployment Q1, Q2 and Q3 were located 10 m below mean tide, close to the bottom, weighed down by 6 kg weights and held upright by a submerged buoy. K1, K2, K3 were deployed similarly but at a shallower depth, 5 m below mean tide. F and G were deployed 5 m below mean tide and held in place by a steel wire attached to a steep cliff, weighed down by 2 kg weight as the steep bottom topography did not allow a similar deployments in these fjord sections.

Fig. 1
figure 1

Map of the Godthåbsfjord system, south west Greenland (64°N, 51°W), outlining the position of stations where settling plates were deployed 2009–2010 (Q1, Q2, Q3) and 2016–2017 (K, F, G), respectively

Full size image

To quantify the physical conditions, HOBO light + temperature loggers (30 min logging rate) were attached directly onto deployments Q2, K2, K3, G and F, and additional HOBO temperature loggers were attached to deployment K1a, K1b, K1c. We obtained surface salinity measurements (10 min logging rate) from moorings at 5 m depth, which were positioned close to the deployment structures for 1 year (Jan ‘16–Jan ‘17) and expected to reflect minimum/maximum values similar to the 1 year deployment period (April ‘16–April ‘17). We calculated day degrees for each station by summing up the mean daily temperatures over a year (April ‘16–April ‘17). To quantify the presence and percent cover of macrobenthos, the seabed was filmed with an underwater camera in April 2017 close to each deployment (0–30 m).

All algal specimens were counted, measured by length, freeze-dried (Christ Alpha 1–2 LD plus) and weighed. For a fraction of the samples that were not freeze-dried, dry weight was estimated based on dw/ww ratios established from the freeze-dried samples. Subsamples for identification of species under microscope were preserved in a solvent of 2/3 corn syrup (Karo light corn syrup), 1/3 water and a few drops of Atamon (own recipe). Mussels and barnacles were identified to species, counted, measured by maximum dimension and freeze-dried. Other invertebrates were counted, but their occurrence was limited. At the underside of the baseplates, we counted and measured barnacles by maximum dimension (i.e. carapace diameter); at high densities, only ½ or ¼ of the plate surface was processed.

To compare the variation in algal biomass per time period between stations (F, G, K) within each of the spatial scales (site, fjord, fjord system), we applied a one-way ANOVA or Kruskal–Wallis test. In the fjord-scale analysis, we compared stations K2, K3 and K1c, the latter chosen randomly among three replicate K1 stations (K1a, K1b, K1c). In the fjord system scale analysis, we compared station G, F and K3, the latter chosen randomly from the five Kobbefjord stations (K1a, K1b, K1c, K2, K3). To assess the relation between barnacle size and day degrees, we applied a linear regression analysis. We used the statistical software R (R Core Team 2017) for analysis and processing of the data.

Results and discussion

Temperature, salinity and light

We found a clear gradient in physical parameters between stations. Salinity and temperature were lowest at the inner part of Godthåbsfjord (G), closest to the glaciers discharging large amounts of freshwater and ice during the melting season, and highest in Kobbefjord (K), furthest away from the glaciers (Fig. 2). At the inner part of Godthåbsfjord (G), the temperature ranged between 0 and 8° in summer and from − 2 to 2° C in winter, displaying the most variable winter temperature compared to all the other stations (Fig. 2). Day degrees, mean daily temperatures accumulated over a year, were highest in Kobbefjord (K) and lowest at the innermost station in Godthåbsfjord (G). Over a year, minimum mean daily salinities varied considerably between stations (Table 1). The lowest mean daily salinity was recorded at the innermost station in Godthåbsfjord (G) (12.4 psu) and the highest (28.3 psu) at the outermost stations [Kobbefjord (K)]. Inconsistency in light logging during 2016–2017 precludes clear patterns in light intensity, but light intensity seemed to be lowest at the inner part of Godthåbsfjord (G) and highest in the outer part (F) (Fig. 2). This trend has also been observed in other studies, suggesting that the high input of silt concomitant with the meltwater strongly impacts the light climate in the fjord system (Murray et al. 2015; Meire et al. 2017).

Fig. 2
figure 2

Water temperature (running mean over 10 observations at a time) and light intensity from station Q2 at 10 m depth 2009–2010 and stations K, F and G at 5 m depth in the Godthåbsfjord system 2016–2017, south west Greenland. In 2016–2017, the light measurements display some inconsistency with sudden drops in the measurements that may be due water movements turning the logger around occasionally

Full size image
Table 1 Position of stations, deployment period and annual minimum/maximum salinity at each of the stations G, F and K
Full size table

Macroalgal recruitment

We found a high density and biomass of settled kelp, Saccharina sp. and Alaria esculenta, at station Q1, while Q2 and Q3 displayed low or no settling of benthos (Online Resource 1). These plates were mainly covered by a thin biofilm. In the Godthåbsfjord, fruiting sori of kelp have been observed in August (Krause-Jensen et al. 2012), suggesting that spores may have settled rather quickly after deployment of plates in August at Q1, and we may have missed the window for settling at the other stations. Saccharina sp. was mostly identified as Saccharina longicruris. Figure 4 illustrates the vertical distribution of blade biomass and density of each kelp species along the upper 5–10 m water column. The highest blade biomass for Saccharina sp. was found on the rope at 6–7 m depth (6800 g dw m−2 year−1); the highest density was on the rope at 5–6 m depth (5866 ind. m−2) (Online Resource 2), and their product, biomass per individual, peaked at 11.6 g ind.−1 year−1 on the buoy at 5 m depth. On the settling plates, blade biomass and density of both kelp species were similar (Saccharina sp.: 977 ind. m−2,520 g dw m−2 year−1; A. esculenta: 647 ind. m−2, 440 g dw m−2 year−1), but low compared to the rope and buoy. The decline in biomass with depth was likely caused by shading and density-dependent mechanisms (Schiel and Foster 2006). Moreover, the different substrate materials may have affected the pattern. While other studies report annual growth of mature kelps in Greenland at up to 175 g dw year−1 and 0.6–2.2 m year−1, with values from the Godthåbsfjord being 130–175 g dw year−1 and 1.8–2 m year−1 (Krause-Jensen et al. 2012), the current findings document that Greenland kelps also have a marked potential for colonization and fast early growth.

Contrary to the earlier observations in 2009–2010, no kelps were found on the settling plates, rope or buoy deployed in 2016–2017, despite kelps being present near the plates. Video-surveying of the surroundings of all deployments in April 2017 (Online Resource 3) showed the presence of kelp but the percent coverage varied greatly, suggesting that also on natural surfaces kelp recruitment may show large variation in time and space. Integrated with the earlier deployments, this shows that kelp recruitment varies greatly from zero settling to 5866 ind. m−2, with an annual production of more than 6000 g dw m−2 year−1. This demonstrates a large growth potential once kelp spores settle, yet recruitment may be a bottleneck in kelp population dynamics. The change in deployment depth and plate material between deployments in 2009–2010 and 2016–2017 was not expected to influence settling of kelp, as settling and high growth also occurred on the buoy at 5 m depth in 2009–2010. Moreover, the plates deployed in 2016–2017 (Online Resource 4) displayed similar biofilm growth as the plates deployed in 2009–2010 at Q2 and Q3 (Online Resource 1). Generally, the recruitment in 2016–2017 was dominated by bushy algae with mean biomasses up to 300 g m−2 (Fig. 3), primarily of the species Desmarestia aculeata in Kobbefjord (K). Other settled algal species were Dictyosiphon foeniculaceus and Ectocarpus sp, the latter dominating in the inner Godthåbsfjord (G).

Fig. 3
figure 3

Algal dry weight (dw) biomass found on settling plates at three spatial scales (site, fjord, fjord system) over three time periods (April–April, April–Aug, Aug–April) within 2016–2017 in the Godthåbsfjord system, south west Greenland (n = 4). Boxplots present the minimum, the lower quartile, the median, the upper quartile and the maximum

Full size image

Settling occurred mainly in the period April–Aug, as we found very little or no settling on plates deployed Aug–Apr (Fig. 3). However, in the inner Godthåbsfjord (G), we found considerable settling by Ectocarpus sp. in Aug–April, and the kelp colonization recorded in 2009–2010 settled in August or later. The algal biomasses developed on the plates from April to August 2017 differed significantly between stations at the site scale, and differences were border-significant (ANOVA, F(2,9) = 4.22, p = 0.051) between stations at the fjord system scale (Table 2). In the period April’16–April ‘17, algal biomass only varied significantly between stations at the fjord system scale (Table 2).

Table 2 One-way ANOVA testing the variation in algal biomass (WW) between stations within a site, fjord or fjord system in the two time periods (April ‘16–Aug ‘16 and April ‘16–April ‘17) (n = 4 at each station per time period). For asteriks symbol marked with a, Kruskal-Wallis test was applied
Full size table

Conclusively, the settlement and growth of macroalgae varied on both the small (20 m) and the large (> 30 km) spatial scale, suggesting that both abiotic and biotic factors may be key drivers of recruitment and early growth. Key physical conditions showed very low variation at the small scale, where biological factors, such as the proximity of the bushy brown algae, D. aculeata, to plates, may have caused the small-scale variation we found. However, abiotic factors unaccounted for, such as flow velocity and local bathymetry, could also be at play. Within the large spatial scale, we found high variation in temperature and salinity (Table 1, Fig. 2) and significant variation in recruitment rates (Table 2), suggesting that temperature and salinity may act as important drivers of recruitment within large spatial scales.

Mussel and barnacle recruitment

We observed high densities of juvenile mussels Mytilus spp. (< 3 mm) entangled in D. aculeata in 2016–2017, but only on plates deployed for the full year, suggesting that settling of Mytilus spp. occurred between August and April likely associated with the presence of D. aculeata. This corresponds to the spawning of Mytilus trossulus in July, observed in a north west Greenlandic fjord (Thyrring et al. 2017), as settlement onto bushy structures is expected 1–2 months after spawning, i.e. Aug/Sep, before moving into the adult patches (Lane et al. 1985). In addition, Gammarus sp. and Caprella septentrionalis were often found entangled in D. aculeata, as also documented in Kongsfjorden, Svalbard (Lippert et al. 2001). However, we also observed Gammarus sp. and C. septentrionalis on settling plates at station F and G, where D. aculeata was limited or absent.

We found that the density of Semibalanus balanoides colonizing the underside of baseplates was quite variable between stations (Q1, Q2 and Q3) in 2010 and the highest density was observed at Q1 (5233 ind.−1 m−2) (Fig. 4a). In August 2017, the density and size of S. balanoides was low on the plates compared to the underside of the baseplate also displaying high variation between sites (Fig. 4b, c). On all 120 plates deployed in 2016–2017, we found 8 individuals of S. balanoides, and none of them were recorded in the period Aug’16–Apr ‘17. This pattern could be related to the fast colonization and growth by D. aculeata on plates posing an overall negative affect on S. balanoides recruitment (Jenkins et al. 1999) or a preference for the material of the baseplates or their position in the water column. Other studies have also reported that recruitment of benthic fauna, including barnacles, is higher on the underside of plates, suggesting that algae are competitively superior on the top side (Khalaman et al. 2016). Hence, modification of the settling surface appears important as indicated by the potential negative effect of algae on S. balanoides and positive effect of algae on Mytilus spp., indicating that timing of settling between different species is decisive in structuring benthic communities.

Fig. 4
figure 4

Semibalanus balanoides after 1 year of colonization on settling plates and/or the underside of baseplates in the Godthåbsfjord system, south west Greenland. a Density at the underside of baseplates after deployment August 2009–August 2010 (n = 1), b density and c size distribution and day degrees on settling plates and the underside of baseplates after deployment April’16–April ‘17 [baseplates: G and F, (n = 1) and K, (n = 2), settling plates G and F, (n = 4) and K, (n = 8)]. Error bars are SE

Full size image

In 2017, the highest density of S. balanoides colonizing undersides of baseplates was found in Kobbefjord (K) (183 ind.−1 m−2) (Fig. 4b), and we already observed colonizers (< 10 mm) 4 months after deployment. This corresponds to earlier documentation of S. balanoides settling occurring in the Arctic mid-summer (Feyling-Hanssen 1953). Both density and growth rate of S. balanoides seemed higher in Kobbefjord (K) than at stations further into the Godthåbsfjord system (F and G) (Fig. 4b, c), yet the low number of replicate baseplates precluded testing for statistical differences. The spatial pattern in S. balanoides recruitment was related to temperature, as day degrees were lower in both stations further into the Godthåbsfjord system (F and G) compared to Kobbefjord stations (Fig. 4c), where also salinity displayed less variation. We found a significant linear relation between S. balanoides size and day degrees (Slope = 0.016, p = 0.007, r2 = 0.912). The mean size of S. balanoides was highest at station K2 in Kobbefjord with 10 mm ind.−1 year−1, and the largest individual was 20 mm. Within Kobbefjord (K), density and growth seemed to vary less compared to stations further into the Godthåbsfjord system, corresponding to a low variation in day degrees (Fig. 4c). In addition to a temperature effect, the large variations in salinity close to the glacier output may also have affected S. balanoides recruitment and growth (Table 1). Earlier work shows that low temperatures in combination with low salinities can impact barnacle recruitment negatively (Nasrolahi et al. 2013).

Conclusively, our results suggest that S. balanoides growth rates and recruitment are affected by significant large-scale patterns in temperature and salinity, created by glacial ice and melt water. Therefore, changes in water temperature and freshening may affect future recruitment patterns and growth rates for marine benthos and ultimately affect benthic community structure. However, we stress that further effort is needed in studying potential drivers of recruitment patterns and growth of Arctic benthic communities, as environmental factors, including the risk of ice scouring, are changing fast.